SPERMATOGENESIS OF THE DRAGON-FLY. 285 



from fresh material in which the loops of the spireme could be 

 made out. Quadripartite bodies similar to those of Sympetrum 

 are .formed from segments of the spireme. At first the longi- 

 tudinal arms of these quadripartite crosses are longer than 

 the transverse arms; but after the formation of the spindle, the 

 four arms are nearly equal in length and the central part is less 

 dense than the arms. 



In a polar view of the primary spermatocyte division, 12 

 chromosomes can always be counted. These are arranged in an 

 irregular ring of 8, surrounding four central chromosomes. More 

 than 200 polar views were observed in which this number was 

 present. Twelve was, therefore, thought to be the reduced 

 number until the behavior of the chromosomes in this division 

 was ascertained. It was then discovered that one chromosome, 

 presumably the sex-chromosome, in nearly all cells goes over in 

 advance of the autosomes to one pole. In consequence of such 

 behavior, a polar view in which this chromosome is visible in 

 the metaphase is difficult to find. Fig. 96, however, shows 13 

 chromosomes in a polar view and this is undoubtedly the correct 

 number for the primary spermatocyte. Fig. 97, taken from a 

 smear, shows only 12 chromosomes. Figs. 98 and 99 are drawn 

 from the same cell at two different focal levels and in 99 the sex- 

 chromosome appears above the autosomes. Figs. 100 to 103 

 inclusive, represent side-views of spindles with the sex-chromo- 

 some in advance of the autosomes. Fig. 104 is from an aceto- 

 carmine smear and does not show the sex-chromosome. Figs. 

 105 to 109 are of especial value as they are drawn from living 

 unstained tissue teased out in salt solution. The chromosomes 

 could be distinguished from the protoplasm and the spindle 

 fibers by the way in which they refracted the light. All stages 

 in division could be found and one cell was watched while it 

 underwent the change from the metaphase to the telophase 

 stage, where further division stopped due probably to the fact 

 that the solution was not absolutely isotonic with the cell proto- 

 plasm. Figs, no, in, and 112 are telophase stages of this 

 division. Fig. 113 is the same stage drawn from a smear prepa- 

 ration. 



In the secondary spermatogonial division two classes of cells 



